I believe that's why the knock sensor is overly conservative. You may never be able to remove all knock events without making the engine run overly rich or overly "retarded". That's one reason why I wish the PV would provide an option to turn the knock sensor off from the PV and without having to pull the map and change it in WINPV and then reload it on the bike. When the knock sensor kicks in, whether it's due to a real ping event or the ion sensor inaccurately recording an event, the performance of the bike is significantly reduced.

If your bike is running well without any audible ping events, try running with good fuel and turning the knock sensor off and you feel the significant difference I'm talking about. In fact if you look at most of the performance maps that come from Harley, I think you'll find that most of them have the knock sensor turned off.

 

It is going to be fairly normal to see variances in the corrections, the key is to getting VE levelized and letting the O2 integrator and the rest of the closed loop system do its work as normal. Regular corrections for fuel quality, conditions, etc.. are all part of the big picture on how the system is designed. Remember, the data you are using with PV is populated by the ECM doing business as usual, any corrections populated in the PV Log tuner software are simply coming from the data bus itself and the ECM is already using these corrections in one way or another.

 

Because the ECM has unique ways with dealing with knock activity it takes a bit of experimentation/testing to get the spark tables optimized, something we spent alot of time on when we were developing our calibrations. It can be pretty easy to chase your tail especially if you are getting events in one place, if you are continually doing log runs and keep allowing PV Log tuner to remove timing you can end up with timing that is way too retarded and performance can suffer greatly. Also note that depending on the specific knock activity the ECM may hold back the timing for several seconds, by that time you may likely be in a different RPM or load area (this is why you may see knock events in the left side of the map when Log tuning) What it comes down to is allowing the system to work as normal, under normal conditions you are still going to see knock events, a sign the system is working as normal. For best overall performance it is crucial to get your spark timing with enough advance where while mild knock activity is normal and the retard corrects effectively, however the system does not need to apply excessive or large corrections that affect performance negatively.

 

Is there a downside to allowing a map that shows only very mild knock to be operated with the knock sensor OFF? Since most of the higher performance maps that come with the SEPST come with the knock sensor turned off, is there a problem turing it off in other maps that are running well and show only very mild knock when the sensor is turned ON.

 

It all really depends on the application and specifics, we do leave the knock retard active in all of our calibrations. For example: on a well tuned bike if you watch knock activity live on your PV display and say you note it pulls 1-2 degrees of timing this is not likely going to affect performance and you should be fine turning it off, however if you are seeing say 5-6 degree events being removed in the same area the adaptive knock retard will adjust the timing table in this area and performance can be affected. The bigger concern would be finding why large amounts of timing are removed in the first place.

 

Thanks for sharing that insight. Do you know if DJ has any plans to enable the Knock sensor to be turned off/on directly from the PV in the future?

 

deleted

 

Jamie, I'd be interested in your opinion on the following situation:

I had my bike dyno tuned and the bike runs incredible. Not a single complaint (other than maybe low mpg). No backfires, no popping on decel, outstanding acceleration, no stumbles anywhere, great low speed maneuverbility, absolutely no pinging. Here's the wot dyno:



I decided to turn on the knock sensor and do some data logging. The map is open loop so the only data worth collecting with my nb sensors is the knock readings. I did my datalogging with identical 93 octane fuel on a similar temp day as the dyno above. Here are the knock readings for the front and rear cyls from my datalogging:






Clearly there's no issue with the rear cyl. Should I just ignore the knock readings in the front cyl? Do you see any issue with just turning off the knock sensor or would you 1) raise the fuel in the higher knock readings or 2) lower the spark advance in the higher knock cells?

 

I would start on the front cylinder with removing 2-3 degrees in the 90 & 100 Kpa columns from 2500-3500 RPM's, run a few logs and see how the rest of the spark table responds. This couple of degrees is not going to affect performance, especially considering what the ECM is currently doing with the knock retard in this area.

 

I assume you meant to say to remove 2-3 degrees from the front cyl (not the rear).